Post-Cardiac Discharge Calls: What We Ask in the First 72 Hours

Most of the complications I worry about after a cardiac procedure don't show up in the cath lab or the recovery ward — they show up on day two, at home, when the patient is alone with a discharge sheet and a phone number they're told to call "if anything feels wrong." The trouble is, patients are terrible judges of what counts as wrong. Breathlessness gets written off as anxiety. Swelling gets written off as normal. By the time someone calls us, it's often day five or six, and what could have been a medication adjustment is now an ER visit.

That's why the first 72 hours after discharge get a structured call from our team, not a leaflet. This is roughly what that call covers, and why each question is there.

The questions that actually matter

• Breathlessness — not "are you breathless," which almost everyone answers no to, but "can you climb the same stairs you could climb before surgery." Comparative questions catch what absolute ones miss.
• Swelling — ankles, and specifically whether it's new or worse than at discharge. New bilateral swelling in week one is one of our highest-yield red flags.
• Wound status — for anyone with a surgical site, whether there's redness, warmth, or drainage, described in the patient's own words rather than a yes/no.
• Medication timing — specifically anticoagulants and beta blockers, because missed or doubled doses in week one are disproportionately common and disproportionately dangerous.
• Weight — a same-scale, same-time-of-day weight check, because a two-kilogram jump in 48 hours is fluid, and fluid after a cardiac event is not something to wait out.

Why this has to be consistent, not occasional

The protocol above isn't complicated. What's hard is doing it for every single patient, every single time, on schedule — not just for the ones who happen to call in worried. A cath lab discharging a dozen patients a week can usually manage that by hand. A cardiology department discharging that many a day cannot, not without either hiring proportionally or accepting that most patients get the leaflet and a prayer.

The protocol was never the hard part. Delivering it identically to every patient, on the exact schedule that matters clinically, is the part that used to fail — quietly, patient by patient.

What changed for us wasn't the questions — we'd had this checklist for years. What changed is that an AI voice agent now runs it on every discharged cardiac patient at the 24, 48, and 72-hour marks, in whichever language the patient is comfortable in, and routes anything that trips a threshold straight to whoever is on call, with the recording attached. The agent doesn't decide what a swollen ankle means. It asks the question the same way every time, and it makes sure the answer reaches a cardiologist within minutes, not whenever the patient decides to call back.

What this looks like when it works

The case that convinced our team wasn't dramatic on the surface: a post-CABG patient answered "a little more tired than usual" on his day-two call — not breathless, not in pain, just tired. That single data point, combined with a two-kilogram weight gain flagged on the same call, was enough for the on-call resident to bring him in that afternoon for early signs of fluid overload. Caught at day two, it was a diuretic adjustment. Left until his week-two follow-up, it likely would have been a readmission.

None of this replaces clinical judgment — it just makes sure the data clinical judgment needs actually reaches someone, on time, for every patient, not only the ones anxious enough to call first.